Implicit conversion from float -13.375796178343947 to int loses precision